Jorge Chacon-Caldera1, Javier Uranga Solchanga1, Paulina Koziol1,2, and Lothar R Schad1
1Computer Assisted Medical Medicine, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany, 2Department of Medical Physics and Biophysics, Faculty of Physics and Applied Computer Science, AGH University of Science and Technology, Krakow, Poland
Synopsis
Prostate MRI is commonly performed using
endorectal coils which are invasive. This is done since body planar arrays are
not sensitive enough for prostate imaging. Increasing sensitivity of an array
for deep structures in the body is not trivial. In this study, we extended the
traditional stacked figure-8 and single loop quadrature pair to add more single
loop coils and enhanced the sensitivity at the depth of the prostate without
increasing the field over a larger lateral area. We compared these arrays to classical
planar approaches and found a factor 1.35 increase in maximum localized |B1-|
using numerical simulations. PURPOSE
Prostate cancer is an important cause of death amongst males. Its
detection using MRI commonly involves the use of endorectal coils which are
invasive to patients. This is necessary to increase the signal-to-noise ratio
(SNR) within the prostate which is deep in the middle of the pelvic cavity. During
signal reception for a given system, patient and protocol, the SNR is
proportional to the absolute reception magnetic field (|B
1-|)
of the array given by a weighted |B
1-| sum of the coil
elements
[1]. |B
1-| has a strong distance dependence from an
evaluated point to a coil element. This distance is increased quickly due to a lateral element placement in planar
arrays which are currently the non-invasive standard in body imaging. In planar arrays, the main SNR gain is achieved
superficially over a large field-of-view. Recent studies have demonstrated
improved SNR by stacking loops and dipoles at 7 T
[2] but dipoles could
have impractical dimensions at 3 T (ideally, length ∝
λ). In this study, the principle of a traditional quadrature stacked figure-8
(F8) and single loop (SL) coil is extended to multiple SL elements. This was
performed to increase |B
1-| at the depth of the prostate.
We compared these arrays to standard planar arrays.
METHODS
Numerical simulations at 123.25 MHz were performed using the finite
difference time domain method (CST MWS, Darmstadt, Germany). Stacks of one F8
and one, two and three SL coils were simulated and compared to planar arrays of
two, three, four, and five SL coils using a cubic phantom of 50x50x50cm³ with ε
= 34 and σ = 0.4 S/m. Figure 1 shows the simulated and compared arrays. In all arrays,
the nearest coil was placed 10 mm above the phantom and additional coils were
placed 1 mm above the nearest neighbor. The overlaps were experimentally
optimized. Coils were simulated using perfect electrical conductors. The wires
were 2 mm in diameter and the tape had 3 mm thickness and 0.1 mm in height. The
elements in planar arrays were arranged in standard patterns and had 32 mm in
radius
[3]. The F8 loops had radii of 65 mm
[4] and the SL had the following
dimensions: 1SL: 80x80 mm², 2SL and 3SL: 50x80 mm². For the evaluation as a
receiving array, the maximum |B
1-| values were evaluated
10 cm inside the phantom. Additionally, the |B
1-| was
normalized to 1 W accepted power to use as figure of merit (|B
1-|
norm).
The maximum coupling between coils was evaluated using the scatter parameters (S-parameters).
RESULTS
All stacked arrays yielded greater values of both |B
1-|
and |B
1-|norm (see Fig. 2). F8+3SL provided
the highest |B
1-| = 3.38e-7 T. This was a factor 1.35
increase compared to 3SL which yielded the highest |B
1-|
of all the planar arrays (2.51e-7 T). The gain in |B
1-|
norm
was a factor 1.11 of the best performing stacked array (2.77e-7 T) compared to 3SL
which yielded 2.49e-7 T under this metric. Furthermore, all the stacked coil
arrays showed better decoupling per number of coils except for F8+3SL (-5.9 dB) compared to 4SL (-9.5
dB). Fig. 3 shows the higher localized |B
1-| yielded by the
stacked arrays.
DISCUSSION
The complexity to increase SNR at a deep point of interest adding coil elements laterally has been demonstrated e.g. in the center of a body sized phantom, a factor 1.03 increase was found when using 128 elements instead of 32
[3]. To our knowledge, more SL elements were added to the classical F8 and SL
quadrature pair for the first time yielding good results for the depth of the
prostate. It was also shown that the |B
1-| is more
focused using the stacked arrays which is beneficial for the small size of the
prostate. All of the tested stacked arrays performed better than 3SL which was the best
planar array in both metrics. It has been previously demonstrated that the quadrature F8
and SL provides higher SNR that 2 SL coils for cardiac imaging
[5]; our results predict the
same pattern for the prostate and additional gains in SNR with added SL coils. Finally, the stacked arrays presented could be used as elements and multiple F8+2SL or F8+3SL arrays
could be built and placed around the body for the use of parallel imaging.
CONCLUSION
The combination of an F8 stacked on 2 and 3 SL coils provided a higher maximum
|B
1-| than standard approaches. This and the possibility
of quadrature detection can considerably increase the local SNR compared to planar
arrays at 3 T for the imaging of the prostate.
Acknowledgements
The first author acknowledges funding from the DAAD Research Scholarship 57076385References
1. Wright SM, et al. Theory and Application of Array Coils in MR Spectroscopy. NRM Biomed. 1997;10:394-410.
2. Voogt IJ, et al. Combined 8-channel transceiver fractionated dipole antenna array with a 16-channel loop coil receive array for body imagingat 7 Tesla. In Proc. ISMRM. 2015;23:0631.
3. Hardy CJ, et al. 128-Channel Body MRI With a Flexible High-Density Receiver-Coil Array. JMRI. 2008; 28:1219-1225.
4. Kumar A, et al. Optimized quadrature surface coil designs. MAGMA. 2008;21:41-52.
5. Ouhlous M, et al. Quadrature Coil Design for High-Resolution Carotid Artery Imaging Scores Better Than a DualPhased-Array Coil Design With the Same Volume Coverage. JMRI. 2007;25:1079-1084.